CN104618924A - Wireless ubiquitous network-based quality of experience index system and measuring method - Google Patents

Abstract

The invention discloses a wireless ubiquitous network-based quality of experience index system and measuring method. The wireless ubiquitous network-based quality of experience index system solves the problem of performance of network, terminals and multimedia services as well as the problem of subjective factors of users. The measuring method comprises, firstly, disclosing an index system according to main factors affecting QoE (quality of experience); then, according to the disclosed index system, establishing an assessment model; lastly, obtaining a comprehensive assessment. The wireless ubiquitous network-based quality of experience index system takes objective factors as well as the subjective factors of the users into consideration, and meanwhile refers to influence of previous experience on current experience, thereby comprehensively and perfectly reflecting the feelings of the users on service experience.

Description

Based on user experience quality index system and the method for measurement of wireless ubiquitous network

Technical field

The present invention relates to a kind of user experience quality index system based on wireless ubiquitous network and method of measurement, belong to wireless communication technology field.

Background technology

In recent years, in order to adapt to different demands, there is different Radio Access Network technology in mobile communication technology develop rapidly, and meanwhile, wireless mobile telecommunication technology is experiencing the evolution of isomery fusion and ubiquitousization.Wireless Ubiquitous Network (WirelessUbiquitous Network), the wireless network namely extensively existed, it is with omnipresent, all-embracing, omnipotent is essential characteristic, to realize any time, any place, anyone, any thing can communicate smoothly as target.Wireless Ubiquitous Network comprises the implication of two aspects, and one is ubiquitous service, and two is for user provides service by wireless access.The final purpose of Ubiquitous Network is for user provides ubiquitous amalgamation service.Ubiquitous Network does not need to rebuild new network, and it is actual is on the basis of legacy network, according to the demand of human social development, increases corresponding network capabilities, service and new application, to enable various resource collaborative and shared fully.

The evolution of wireless network is more perfect in order to be to provide user, better service, so, user will become the core index evaluating wireless network quality gradually to the experience of QoS, a kind of concept of novelty is arisen at the historic moment in this context---user experience quality (Quality of Experience, QoE).QoE is a kind of with the evaluation method of the customer acceptance degree service that is standard, it combines the influencing factor of service aspect, user level, environment aspect, directly reflects the degree of recognition of user to service.International Telecommunication Union is defined as QoE: terminal use is to application or serve overall subjective acceptable degree.

The feature of the mobile and network of mobile interchange net collective, provides to user and more networks convenient, and create data service miscellaneous.Traditional language service is saturated, and mobile data services will be new profit growth points.Along with continuing to bring out of data service, network service is constantly to customization future development, and user is improving constantly the quality of business and individual requirement.Wireless at ambient general, multimode terminal can access different networks, therefore, how to realize effective utilization of Internet resources between different radio Access Network and terminal capability, improve formation transfer service quality, from different perspectives, different levels meet consumers' demand, thus improve user experience quality QoE and seem very important.And the present invention can solve problem above well.

Summary of the invention

The object of the invention is to propose a kind of user experience quality index system based on wireless ubiquitous network and method of measurement, and this system and method for measurement are all based on wireless network, evaluate mobile data services in conjunction with fuzzy mathematics and analytic hierarchy process (AHP).The invention solves and affect the subjective factor of user experience quality and the problem of objective factor, set up index system and the assessment models of a layering, to present the overall merit of user to business well.The weight of each factor in application Fuzzy AHP parameter system, and finally draw a comprehensive evaluation score, comparatively intuitively and truly react the impression of user.Meanwhile, the method is simple to operate and be easy to realize, and has good application prospect.

The present invention solves the technical scheme that its technical problem takes: the invention provides a kind of user experience quality index system based on wireless ubiquitous network and method of measurement, the method not only solves the problem of the performance of network performance, terminal capabilities and multimedia service, but also solves user's subjective factor problem.First, propose an index system according to the principal element affecting QoE, then according to proposed index system, set up an assessment models, finally draw an overall merit.

User experience quality of the present invention (that is: QoE) is a kind of with the evaluation method of the customer acceptance degree service that is standard, and it combines objective factor and subjective factor, directly reflects the degree of recognition of user to service.Wherein objective factor comprises the performance of terminal capabilities, network performance and business itself, and subjective factor comprises mood, user's expectation of user, self background and environmental factor.

System of the present invention comprises two main modular compositions, that is: objective factor module and subjective factor module.

The major function of objective factor module objective factor is quantized to evaluate the experience sense of user to business to the impact of QoE by calculating be subject to, wherein comprise again three submodules: (1) terminal capabilities module QoT (Quality of Terminal), calculate the CPU of mobile terminal, fortune deposits RAM and screen resolution to the impact of user experience quality, and obtain evaluation of estimate; (2) network performance module QoN (Quality of Network), calculate the packet loss of wireless network, shake, time delay and bandwidth on the impact of user experience quality, and obtain evaluation of estimate; (3) service feature module QoA (Quality of Application), calculates the instantaneity of mobile data services, accessibility, content quality, operability and fail safe to the impact of user experience quality, and obtains evaluation of estimate.Finally by three submodule gained to quantized value obtain a comprehensive objective factor evaluation of estimate by weighted calculation.

The major function of subjective factor module is by calculating an evaluation of estimate reflection active user to the impression of business the subjective factor of user self, represent with IoS (Indicator of Subjectivity), its influencing factor has user to expect, mood, self background, environment.

Can the experience sense of concentrated expression user to business be subject to by the evaluation of estimate of weighted calculation two main functional modules.

Method flow:

Step 1: build network environment, under real wireless Ubiquitous Network environment, tests, and the parameter in adjustment performance indicator layer, obtains experimental data and arrange, selecting corresponding data model to evaluate according to KQIs layer with the mapping relations of KPIs layer.

Step 2; Applied statistics regression model evaluating network performance and terminal capabilities, experimentally data analysis, use regression analysis method, obtain the high regression equation of degree of fitting for quantizing terminal capabilities and network performance to the impact of Consumer's Experience, if regression equation carries out verification degree of fitting height, regression equation is, otherwise repeat step 2 until the degree of fitting of regression equation meets the requirements, regression equation is as follows:

MOS _QoT＝a*CPU ^b+c*RAM ^d+e*SR ^f+g

${\mathrm{MOS}}_{\mathrm{QoN}}=\frac{a+b*\mathrm{In}\left(\mathrm{BW}\right)}{c+(d*\mathrm{Delay}+e*\mathrm{Jitter})*e^{\mathrm{PKL}}}$

Wherein MOS _qoT, MOS _qoNthe evaluation of estimate of terminal capabilities QoT and network performance QoN respectively;

Step 3: application fuzzy mathematical model evaluation assignment performance and customer impact index.Survey is carried out to the people of all ages and classes, sex, educational background, user is allowed to compare the significance level between two between factor to the influencing factor of five in service feature QoA, ultimate analysis data, obtain the fuzzy judgment matrix of service feature QoA five factors, in the same way, obtain the fuzzy judgment matrix of four factors of user's subjective impact, application Fuzzy AHP, obtain the weight vectors of influencing factor and obtain quantification equation:

MOS _QoA＝w ₁*mos ₁+w ₂*mos ₂+…+w ₅*mos ₅

${\mathrm{MOS}}_{\mathrm{IoS}}={\stackrel{\‾}{w}}_1*{\stackrel{\‾}{\mathrm{mos}}}_1+{\stackrel{\‾}{w}}_2*{\stackrel{\‾}{\mathrm{mos}}}_2+...+{\stackrel{\‾}{w}}_4*{\stackrel{\‾}{\mathrm{mos}}}_4$

Step 4:: obtaining the evaluation of estimate MOS of four factor network performance QoN, terminal capabilities QoT, service feature QoA and user's subjective factors in quality index layer KPIs _qoN, MOS _qoT, MOS _qoA, MOS _ioSbasis on, by calculate obtain this user to the evaluation of estimate MOS of business experience quality _qoEN, formula is as follows:

MOS _QoEN＝W _QoT*MOS _QoT+W _QoN*MOS _QoN+W _QoA*MOS _QoA+W _IoS*MOS _IoS

Step 5: the evaluation of estimate MOS experienced according to this _qoENwith the evaluation of estimate MOS that last time is experienced _qoEPbetween relation, calculate according to psychologic fuzzy rule, finally obtain one rationally, exactly reflect user to the comprehensive evaluation value of the sense of reality of business:

MOS _QoE＝(1-p _ij)*MOS _QoEP+p _ij*MOS _QoEN。

Beneficial effect:

1, the user experience quality index system that the present invention proposes not only considers objective factor, and considers the subjective factor of user, experiences the impact experienced this simultaneously, fully reacted the impression of user to business experience perfectly with reference to last time.

2, applied regression analysis of the present invention and Fuzzy AHP, quantizes to obtain comprehensive evaluation value by user experience quality by founding mathematical models.

3, the index system that proposes of the present invention and evaluation method simple, be easy to realize, there is good application prospect.

Accompanying drawing explanation

Fig. 1 is QoE index system structure chart of the present invention.

Fig. 2 is method flow diagram of the present invention.

Embodiment

Below in conjunction with Figure of description, the invention is described in further detail.

As shown in Figure 1, the present invention obtains the QoE index system of a tree-like hierarchy, whole index system is a tree structure of standing upside down, there is a unique root node and user experience quality, according to certain level expansion branch and leaf nodes, whole index system has strict hierarchical structure, comprises altogether three layers: experience layer, quality index layer (that is: KQIs) and performance index layer (that is: KPIs).The index system of layering can reflect mapping relations between layers, and the subjectivity of people is separated with objectivity index, can apply better and quantitatively evaluate with the Quality of experience of the qualitative method combined to user.

Analytic hierarchy process (AHP) (that is: AHP) does not consider the judgement ambiguity of people when Judgement Matricies, simultaneously too complicated when consistency check, practical level is not high.Regression analysis (regression analysis) is a kind of statistical analysis technique determining complementary quantitative relationship between two or more variable, uses very extensive, is generally used for data analysis.Based on the index system of the QoE tree-like hierarchy of above-mentioned proposition, Fuzzy AHP (Fuzzy Analytical HierarchyProcess, FAHP) and regression analysis method is used to be quantized by QoE.

More widely used QoE quantization method is that " average point value of evaluation " (MeanOpinion Score advises in national telecommunication alliance (ITU) at present, MOS), the subjective feeling of people is divided into 5 levels by the method, as shown in table 1, the experience sense that this quantization method describes user is meticulously subject to.

MOS	QoE	The extent of damage
			5	Excellent	Can not discover
4	Good	Discernable but not serious
			3	In	Slightly
2	Secondary	Seriously
			1	Bad	Very serious

The average point value of evaluation (MOS) of table 1.

The terminal capabilities (QoT) of quality index layer is affected by three factors, application Multiple Non-linear Regression Analysis, obtains the quantized value of QoT, determines that dependent variable is the quantized value of QoT, independent variable is the value of CPU handling property, internal memory performance and screen resolution, obtains regression equation:

MOS _QoT＝a*CPU ^b+c*RAM ^d+e*SR ^f+g (1)

Wherein MOS _qoTrepresent the quantized value of terminal capabilities, value scope is 1 to 5; CPU represents that the dominant frequency size (GHz) of terminal CPU reflects the handling property of CPU; Ram table shows that terminal is transported the size (G) of depositing and reflected the handling property transported and deposit; SR represents that terminal screen resolution size (ten million pixel) reflects the definition of terminal screen.

For ensureing the accuracy of regression equation, needing to verify it, mainly adopting standard deviation (σ), goodness of fit verification (coefficient of determination R ²verification) and the checking validity (F verification) of regression equation.Wherein standard deviation (σ) reflects the departure degree of user experience value and experimental data, and be worth less, departure degree is less, and formula is as follows:

$\mathrm{\σ}=\sqrt{\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{({\hat{y}}_i-y_i)}^2}---\left(2\right)$

Wherein, represent the MOS value that regression fit obtains, y _irepresent the value in experimental data.

Goodness of fit verification (coefficient of determination R ²verification) reflect the fitting degree of regression equation, R ²more close to 1, the degree of fitting of regression equation is better, and formula is as follows:

$R^2=1-\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{({\hat{y}}_i-y_i)}^2}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{({\stackrel{\‾}{y}}_i-y_i)}^2}---\left(3\right)$

Wherein, represent the mean value of MOS in experimental data.

F distribution is defined as follows:

$F=\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{({\hat{y}}_i-{\stackrel{\‾}{y}}_i)}^2}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{({\hat{y}}_i-y_i)}^2}*\frac{n-m-1}{m}~F(m,n-m-1)---\left(4\right)$

X={F > F _α(m, n-m-1) }, the value of F is more greater than critical region, and then regression effect is better.Wherein m is independent variable number, and n is observation group's number, and α is given confidence level critical region.

By obtaining a after verification, the parameter value of b, c, d, e, f, g obtains the good regression equation of a degree of fitting, for quantizing terminal capabilities.

Network performance (QoN) is affected by 4 factors, applies Multiple Non-linear Regression Analysis equally, obtains the quantized value of QoN, determines that dependent variable is the quantized value of QoN, and independent variable is the value of packet loss, shake, time delay and bandwidth, obtains regression equation:

${\mathrm{MOS}}_{\mathrm{QoN}}=\frac{a+b*\mathrm{In}\left(\mathrm{BW}\right)}{c+(d*\mathrm{Delay}+e*\mathrm{Jitter})*e^{\mathrm{PKL}}}---\left(5\right)$

Wherein MOS _qoNrepresent the quantized value of network performance QoN, scope is 0 to 5; BW represents bandwidth (M); Delay represents time delay (ms); Jitter represents shake (ms); PKL represents packet loss (%).Obtain a good regression equation of degree of fitting by relevant verification equally and quantize network performance.

In objective factor, in quality index layer, the 3rd factor service feature (QoA) reflects the correlated performance of business, and the total class of mobile data services is various, has the instant messaging of communication class, mail etc.; There are the news category information of information content class, weather class, position class information etc.; There is the mobile payment of transaction class, stored value card etc.Extract five main performance index and quantize service feature (QoA) as a reference, due to the ambiguity between index, use Fuzzy AHP (FAHP) to quantize service feature.

Matrix R=(r _ij) _n*nmeet 0≤r _ij≤ 1; I=1,2 ..., n; J=1,2 ..., n then claims R to be fuzzy matrix; If meet r _ij+ r _ji=1, then claim R to be fuzzy complementary matrix; If meet r _ij=r _ik-r _jk+ 0.5 (k=1,2 ..., n), then claim R to be Fuzzy consistent matrix.

Evaluate element of set element a ₁, a ₂..., a _nimportant degree obtains fuzzy matrix R and is between two

$R=\left(\begin{array}{cccc}{r}_{11}& r_{12}& ...& r_{1n}\\ r_{21}& r_{22}& ...& r_{2n}\\ ...& ...& ...& ...\\ r_{n1}& r_{n2}& ...& r_{\mathrm{nn}}\end{array}\right)---\left(6\right)$

Evaluate element a ₁, a ₂..., a _nweight be respectively w ₁, w ₂..., w _n, wherein r _ijrepresent a _icompare a _jimportant degree of membership, r _ijlarger expression a _icompare a _jmore important; r _ijless expression two element significance levels are the same.In addition, weighted value w _irepresent evaluation element a _ithe one tolerance of significance level, w _ilarger then a _imore important.Evaluate element a _iand a _jdegree of membership r _ijcan by w _i-w _jfunction represent, defined function f is as follows:

r _ij＝f(w _i-w _j),-1≤w _i-w _j≤1 (7)

By weierstrass (Weirst rass) theorem, polynomial computation is carried out to function f, and according to the relevant nature of Fuzzy consistent matrix, the concrete form finally obtaining function is as follows:

r _ij＝0.5+a*(w _i-w _j),0≤a≤0.5；i＝1,2,…,n；j＝1,2,…,n (8)

Wherein a represents and measures the one of the subjective differences degree of investigated object, can select a satisfied weight vectors by adjustment a.When fuzzy matrix R is inconsistent time, adopt least square method to solve constraint planning problem and obtain weight vectors W=(w ₁, w ₂..., w _n) ^t.

$\left\{\begin{array}{c}\min z=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{[0.5+a*(w_i-w_j)-r_{\mathrm{ij}}]}^2\\ s.t.\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{w}_i=1,w_i>0,(i\≤i\≤n)\end{array}\right.---\left(9\right)$

By method of Lagrange multipliers, introduce Laggrange multiplier λ and formula (9) be equivalent to following formula:

$\min L(w,\mathrm{\λ})=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{[0.5+a*(w_i-w_j)-r_{\mathrm{ij}}]}^2+2\mathrm{\λ}*(\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{w}_i-1)---\left(10\right)$

L (w, λ) is asked about w _i(i=1,2 ..., partial derivative n) also makes formula be zero obtain n equation group, adds constraints w ₁+ w ₂+ ... + w _n=1, can solve and obtain weight vectors W=(w ₁, w ₂..., w _n) ^t.

Service feature (QoA) has five influencing factors, then set up factor of evaluation collection (a ₁, a ₂..., a ₅), a ₁represent instantaneity; Represent accessibility; Represent content quality; Represent operability; Represent fail safe, the span of the value that factor of evaluation is concentrated is 1 to 5, represents the quantized value of each influencing factor, constructs fuzzy fuzzy judgment matrix A to be simultaneously:

$A=\left(\begin{array}{cccc}{r}_{11}& r_{12}& ...& r_{15}\\ r_{21}& r_{22}& ...& r_{25}\\ ...& ...& ...& ...\\ r_{51}& r_{52}& ...& r_{55}\end{array}\right)---\left(11\right)$

In order to allow degree of membership r _ijquantized, adopt the 0.1-0.9 quantity scale method generally used to quantize degree of membership.Being described below of this scaling law:

Table 2 represents: 0.1-0.9 quantity scale

By above-mentioned scaling law Full Fuzzy judgment matrix A, solve by formula (9) and formula (10) the weight vectors W obtaining service feature (QoA) by least square method _qoA=(w ₁, w ₂..., w ₅) ^t, also can by fuzzy judgment matrix A according to the character of Fuzzy consistent matrix by certain fuzzy consistent judgment matrix B that calculates be first:

$A=\left(\begin{array}{cccc}\stackrel{\‾}{r_{11}}& \stackrel{\‾}{r_{12}}& ...& \stackrel{\‾}{r_{15}}\\ \stackrel{\‾}{r_{21}}& \stackrel{\‾}{r_{22}}& ...& \stackrel{\‾}{r_{25}}\\ ...& ...& ...& ...\\ \stackrel{\‾}{r_{51}}& \stackrel{\‾}{r_{52}}& ...& \stackrel{\‾}{r_{55}}\end{array}\right)---\left(12\right)$

Calculate weight vectors by formula (8) again, by the adjustment to a, finally draw the weight vectors W of a satisfaction _qoA=(w ₁, w ₂..., w ₅) ^t, then corresponding according to factor of evaluation quantized value, can obtain the final quantization value of service feature (QoA):

MOS _QoA＝w ₁*mos ₁+w ₂*mos ₂+…+w ₅*mos ₅(13)

Wherein mos ₁represent the quantized value of instantaneity; mos ₂represent the quantized value of accessibility; mos ₃represent the quantized value of content quality; mos ₄represent the quantized value of operability; mos ₅represent the quantized value of fail safe, w ₁, w ₂..., w ₅represent the weighted value of corresponding factor successively.

In quality index layer, 4 influencing factors of subjective factor (IOS) directly affect the perception to business of user from the angle of user, then there is fuzzy relation again between these 4 influencing factors, use Fuzzy Level Analytic Approach (FAHP) method the impact of subjective factor to be quantized equally.The fuzzy matrix of the factor of evaluation of subjective factor (IOS) is constructed by said method:

$R_{\mathrm{IoS}}=\left(\begin{array}{cccc}\stackrel{\‾}{r_{11}}& \stackrel{\‾}{r_{12}}& ...& \stackrel{\‾}{r_{14}}\\ \stackrel{\‾}{r_{21}}& \stackrel{\‾}{r_{22}}& ...& \stackrel{\‾}{r_{24}}\\ ...& ...& ...& ...\\ \stackrel{\‾}{r_{41}}& \stackrel{\‾}{r_{42}}& ...& \stackrel{\‾}{r_{44}}\end{array}\right)---\left(14\right)$

Then the weight vectors value of 4 factors of evaluation of subjective factor (IoS) is finally obtained according to formula (8) or formula (10) quantized by subjective factor (IoS), formula is as follows:

${\mathrm{MOS}}_{\mathrm{IoS}}={\stackrel{\‾}{w}}_1*{\stackrel{\‾}{\mathrm{mos}}}_1+{\stackrel{\‾}{w}}_2*{\stackrel{\‾}{\mathrm{mos}}}_2+...+{\stackrel{\‾}{w}}_4*{\stackrel{\‾}{\mathrm{mos}}}_4---\left(15\right)$

Wherein represent the quantized value that user expects; represent the quantized value of mood; represent the quantized value of self background; represent the quantized value of environmental factor, the span of quantized value is all 0 to 5, represent the weighted value of above 4 factors of evaluation respectively.So just can obtain the quantized value MOS of subjective factor _ioS, span is the number of 0 to 5.

So far, obtain the mapping relations between quality index layer and performance index layer and four of quality index layer factors have been quantized, the mapping relations experiencing this experience of layer and 4 factors of quality index layer determined by following needs, and this experience experienced is being quantized.A complete user experience quality not only will consider objective evaluation, and also will consider the subjective feeling of user, the experience sense of user to business is subject to only have both combinations to reflect more accurately.Meanwhile, after combining subjectivity and objectivity factor, there is ambiguity again between 4 influencing factors of quality index layer, Fuzzy AHP (FAHP) combines the method for quantitative and qualitative analysis, can guarantee the reasonability of model.This experiences and quantizes by the above-mentioned FAHP method of same application, finally obtains quantitative formula:

MOS _QoEN＝W _QoT*MOS _QoT+W _QoN*MOS _QoN+W _QoA*MOS _QoA+W _IoS*MOS _IoS(16)

Wherein MOS _qoENrepresent the quantized value of this Consumer's Experience, span is the weight vectors W of 0 to 5,4 factors of evaluation _qoEN=(W _qoT, W _qoN, W _qoA, W _ioS) ^trepresent the weighted value of corresponding influencing factor, so just can obtain the quantized value that this user experiences business experience.

From psychologic angle, the previous experience experience of user also will have influence on this experience sense of user and be subject to, application fuzzy theory, set up a kind of fuzzy relation matrix to reflect previous experience and this experience direct fuzzy relation.The scope of the quantification value of Consumer's Experience is 0 to 5, and layering five grades are respectively 1,2,3,4,5, span is respectively (0,1], (1,2], (2,3], (3,4], (4,5], MOS _qoEPrepresent the quantized value that last time is experienced, MOS _qoEPrepresent this quantized value experienced, two values are bound in the scope of a certain grade in five grades, meanwhile, and MOS _qoEPand MOS _qoENpresent grade is different, and the fuzzy relation between them, also by difference, first builds a fuzzy relation matrix from learning angle at heart:

$P=\left(\begin{array}{cccc}{p}_{11}& p_{12}& ...& p_{15}\\ p_{21}& p_{22}& ...& p_{25}\\ ...& ...& ...& ...\\ p_{51}& p_{52}& ...& p_{55}\end{array}\right)---\left(17\right)$

Wherein, p _ijthe grade of expression experience last time is i and this grade experienced is the direct fuzzy relation of j, and span is 0≤p _ij≤ 1, this fuzzy relation is quantized, obtains one based on psychologic fuzzy rule, as follows:

Table 3. fuzzy rule

Based on above-mentioned fuzzy relation, finally can obtain the comprehensive evaluation value of user to business experience quality, formula is as follows:

$\left\{\begin{array}{c}{\mathrm{MOS}}_{\mathrm{QoT}}=a*{\mathrm{CPU}}^b+c*{\mathrm{RAM}}^d+e*{\mathrm{SR}}^f+g\\ {\mathrm{MOS}}_{\mathrm{QoN}}=\frac{a+b*\mathrm{In}\left(\mathrm{BW}\right)}{c+(d*\mathrm{Delay}+e*\mathrm{Jitter})*e^{\mathrm{PKL}}}\\ {\mathrm{MOS}}_{\mathrm{QoA}}=w_1*{\mathrm{mos}}_1+w_2*{\mathrm{mos}}_2+...+w_5*{\mathrm{mos}}_5\\ {\mathrm{MOS}}_{\mathrm{IoS}}={\stackrel{\‾}{w}}_1*{\stackrel{\‾}{\mathrm{mos}}}_1+{\stackrel{\‾}{w}}_2*{\stackrel{\‾}{\mathrm{mos}}}_2+...+{\stackrel{\‾}{w}}_4*{\stackrel{\‾}{\mathrm{mos}}}_4\\ {\mathrm{MOS}}_{\mathrm{QoEN}}=W_{\mathrm{QoT}}*{\mathrm{MOS}}_{\mathrm{QoT}}+W_{\mathrm{QoN}}*{\mathrm{MOS}}_{\mathrm{QoN}}+W_{\mathrm{QoA}}*{\mathrm{MOS}}_{\mathrm{QoA}}+W_{\mathrm{IoS}}*{\mathrm{MOS}}_{\mathrm{IoS}}\\ {\mathrm{MOS}}_{\mathrm{QoE}}=(1-p_{\mathrm{ij}})*{\mathrm{MOS}}_{\mathrm{QoEP}}+p_{\mathrm{ij}}*{\mathrm{MOS}}_{\mathrm{QoEN}}\end{array}\right.$

According to said method, by Nonlinear Multiple Regressive Analysis method and Fuzzy AHP, in conjunction with subjectivity and objectivity factor, draw the comprehensive evaluation value MOS of a business _qoE, more rationally, more perfect, more can embody the experience sense of user to business and be subject to.

The user experience quality index system of the tree-like hierarchy that the present invention proposes, combines user level, technological layer and service layer, takes into full account the subjective factor of user, reflect the authentic assessment of user to business more rationally, more really.

The method of the present invention's objective factor certain applications multiple regression analysis in index system, test under real scene, obtain the regression equation that degree of fitting is high, be mapped as the function model of quality index as performance index, thus obtain a reasonable correct evaluation of estimate.

The impact of subjective factor on QoE of user of the present invention is difficult to quantize, between the influence index of subjective factor, relation is difficult to determine, application Fuzzy AHP (FAHP) overcomes the limitation of human thinking's subjectivity, process is difficult to the challenge with quantitative assay effectively, subjective factor is quantized the impact of QoE, and in conjunction with the evaluation of estimate of objective factor, the final comprehensive evaluation value obtaining user and business experience is experienced.

As shown in Figure 2, the method for measurement that the present invention is based on the user experience quality index system of wireless ubiquitous network comprises the steps:

The first step: build wireless network environment, the value of the influencing factor of adjustment performance indicator layer, allows multidigit tester carry out subjective assessment, obtains complete experimental data.

Second step: organize experimental data for obtaining, applied regression analysis method, obtains the regression equation that degree of fitting is high, determines the parameter of formula (1) and formula (5) more.

3rd step: for the impact of subjective factor, adopts the form of survey, finally carries out analyzing the weighted value that the fuzzy judgment matrix R obtained in Fuzzy AHP is used for determining correlative factor.

4th step: integrated by the function model of structure, finally obtains the comprehensive evaluation value MOS of a user experience quality _qoEequation, namely can be applicable to the evaluation that practical matter carries out QoE.

Above a kind of QoE evaluation method provided in the invention process is described in detail, for one of ordinary skill in the art, according to the thought of the embodiment of the present invention, all will change in specific embodiments and applications, in sum, the embodiment of the present invention should not be construed as limitation of the present invention.

Claims (5)

1. based on a method of measurement for the user experience quality index system of wireless ubiquitous network, it is characterized in that: described method comprises the steps:

Step 1: build network environment, under real wireless Ubiquitous Network environment, tests, and the parameter in adjustment performance indicator layer, obtains experimental data and arrange, selecting corresponding data model to evaluate according to KQIs layer with the mapping relations of KPIs layer;

Step 2; Applied statistics regression model evaluating network performance and terminal capabilities, experimentally data analysis, use regression analysis method, obtain the high regression equation of degree of fitting for quantizing terminal capabilities and network performance to the impact of Consumer's Experience, if regression equation carries out verification degree of fitting height, regression equation is, otherwise repeat step 2 until the degree of fitting of regression equation meets the requirements, regression equation is as follows:

MOS _QoT＝a*CPU ^b+c*RAM ^d+e*SR ^f+g

${\mathrm{MOS}}_{\mathrm{QoN}}=\frac{a+b*\mathrm{In}\left(\mathrm{BW}\right)}{c+(d*\mathrm{Delay}+e*\mathrm{Jitter})*e^{\mathrm{PKL}}}$

Wherein MOS _qoT, MOS _qoNthe evaluation of estimate of terminal capabilities QoT and network performance QoN respectively;

Step 3: application fuzzy mathematical model evaluation assignment performance and customer impact index;

Survey is carried out to the people of all ages and classes, sex, educational background, user is allowed to compare the significance level between two between factor to the influencing factor of five in service feature QoA, ultimate analysis data, obtain the fuzzy judgment matrix of service feature QoA five factors, in the same way, obtain the fuzzy judgment matrix of four factors of user's subjective impact, application Fuzzy AHP, obtain the weight vectors of influencing factor and obtain quantification equation:

MOS _QoA＝w ₁*mos ₁+w ₂*mos ₂+…+w ₅*mos ₅

${\mathrm{MOS}}_{\mathrm{IoS}}={\stackrel{\‾}{w}}_1*{\stackrel{\‾}{\mathrm{mos}}}_1+{\stackrel{\‾}{w}}_2*{\stackrel{\‾}{\mathrm{mos}}}_2+...+{\stackrel{\‾}{w}}_4*{\stackrel{\‾}{\mathrm{mos}}}_4$

Step 4:: obtaining the evaluation of estimate MOS of four factor network performance QoN, terminal capabilities QoT, service feature QoA and user's subjective factors in quality index layer KPIs _qoN, MOS _qoT, MOS _qoA, MOS _ioSbasis on, by calculate obtain this user to the evaluation of estimate MOS of business experience quality _qoEN, formula is as follows:

MOS _QoEN＝W _QoT*MOS _QoT+W _QoN*MOS _QoN+W _QoA*MOS _QoA+W _IoS*MOS _IoS

Step 5: the evaluation of estimate MOS experienced according to this _qoENwith the evaluation of estimate MOS that last time is experienced _qoEPbetween relation, calculate according to psychologic fuzzy rule, finally obtain one rationally, exactly reflect user to the comprehensive evaluation value of the sense of reality of business:

MOS _QoE＝(1-p _ij)*MOS _QoEP+p _ij*MOS _QoEN

。

2. the method for measurement of a kind of user experience quality index system based on wireless ubiquitous network according to claim 1, it is characterized in that: described method is the environment based on wireless ubiquitous network, the subjective factor of analysis mobile data services and objective factor are on the impact of user experience quality, according to two major influence factors, index system is divided into two main modular, namely objective factor module is used for calculating objective factor and is used for calculating the evaluation of estimate of user's subjective factor impact to business experience on the evaluation of estimate of QoE and subjective factor module, wherein objective factor module contains again three submodules, terminal capabilities respectively, network performance and service feature, three submodules are respectively by corresponding calculated with mathematical model, analyze the influencing factor of each submodule, obtain evaluation of estimate and the MOS of a reflection submodule performance _qoT, MOS _qoN, MOS _qoA, same subjective factor module calculates the evaluation of estimate MOS of user's Subjective Factors to user's factor index (IoS) _ioS.

3. based on a user experience quality index system for wireless ubiquitous network, it is characterized in that, described system comprises objective factor module and subjective factor module;

The function of objective factor module objective factor is quantized to evaluate the experience sense of user to business to the impact of QoE by calculating be subject to, wherein comprise again three submodules: (1) terminal capabilities module QoT, calculate the CPU of mobile terminal, fortune deposits RAM and screen resolution to the impact of user experience quality, and obtain evaluation of estimate; (2) network performance module QoN, calculate the packet loss of wireless network, shake, time delay and bandwidth on the impact of user experience quality, and obtain evaluation of estimate; (3) service feature module QoA, calculate the instantaneity of mobile data services, accessibility, content quality, operability and fail safe to the impact of user experience quality, and obtain evaluation of estimate, finally by three submodule gained to quantized value obtain a comprehensive objective factor evaluation of estimate by weighted calculation;

The function of subjective factor module be the subjective factor of user self by calculating an evaluation of estimate reflection active user to the impression of business, represent with IoS, its influencing factor have user to expect, mood, self background, environment; Can the experience sense of concentrated expression user to business be subject to by the evaluation of estimate of weighted calculation two main functional modules.

4. a kind of user experience quality index system based on wireless ubiquitous network according to claim 1, it is characterized in that: described system is by network performance QoN according to Key Performance Indicator and Key Quality Indicator, terminal capabilities QoT, service feature QoA and user's subjective factor (IoS) divide at quality index layer KPIs, by QoT, QoA, the influencing factor that QoN and IoS is corresponding is divided at performance index layer KQIs, analyze the relation that this is experienced and last time is experienced simultaneously, obtain the index system of three layers, be once from top to bottom, performance index layer, quality index layer and experience layer.

5. a kind of user experience quality index system based on wireless ubiquitous network according to claim 1, it is characterized in that: the user experience quality of described system combines objective factor and subjective factor, directly reflect the degree of recognition of user to service, wherein objective factor comprises the performance of terminal capabilities, network performance and business itself; Subjective factor comprises mood, user's expectation of user, self background and environmental factor.